Speaker
Description
In this article, we propose an approach to ensure the confidentiality of data from data warehouses in a cloud computing context.
Summary
Data warehouses are widely used in the fields of Big Data and Business Intelligence. Their use allows you to have multidimensional and up-to-date views of a company's activity. The relocation of these environments to the cloud is a real trust issue. Companies that find themselves in such a context want to ensure the best possible security of their data since they are outsourced. The existing data security solutions are grouped into 3 areas: confidentiality, integrity and availability. We propose to ensure the confidentiality of data in a context as presented from access control. Several studies and analyses confirm our choice of the RBAC (Role Based Access Control) mechanism among many others to ensure access control to a data. RBAC is a security policy based on the roles assigned to a given user. Data permissions are assigned to roles. Then these roles are assigned to the users. It happens that in a company or organization, we are confronted with a hierarchy of roles. This phenomenon leads to unauthorized access to certain data using joins. Several studies exist on the inferences of roles and permissions assigned to users. Some proposed a topological sorting to prune the graph to eliminate removable concepts. Others proposed an automatic algorithm that uses constraints to filter inconsistent situations in assigning permissions and roles to users. Our approach consists in proposing a means of granting permission that predicts bad inferences and proposes a matrix of possible and optimal pairs (permissions, roles) for which we will have no case of unauthorized access to a resource.
Keywords: Data warehousing, Confidentiality, Cloud computing, Inferences, Prediction
